Beyond Scale: the Diversity Coefficient as a Data Quality Metric Demonstrates LLMs are Pre-trained on Formally Diverse Data

Brando Miranda, Alycia Lee, Sudharsan Sundar, Sanmi Koyejo

Primary Area: generative models

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Keywords: machine learning, large language models, metrics, data diversity, data, data quality

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2024/AuthorGuide.

TL;DR: We rigorously test a formal quantitive metric for data diversity for LLMs and conjecture it helps downstream performance

Abstract: Current trends to pre-train capable Large Language Models (LLMs) mostly focus on scaling of model and dataset size. However, the of pre-training data is an important factor for training powerful LLMs, yet it is a nebulous concept that has not been fully characterized. Therefore, we use the recently proposed Task2Vec diversity coefficient to understand formal aspects of data \textit{quality} that go beyond scale alone. Specifically, we measure the diversity coefficient of publicly available pre-training datasets to demonstrate that their formal diversity is high when compared to theoretical lower and upper bounds. In addition, to build confidence in the diversity coefficient, we conduct interpretability experiments and find that the coefficient aligns with intuitive properties of diversity, e.g., it increases as the number of latent concepts increases. We conclude the diversity coefficient is reliable and conjecture it can be used to build useful diverse datasets for LLMs.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors' identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 6529